hotfix #206

bayesflow/networks/deep_set/invariant_module.py

@@ -58,7 +58,7 @@ def __init__(
         # Outer fully connected net for sum decomposition: inner( pooling( inner(set) ) )
         self.outer_fc = keras.Sequential(name="InvariantOuterFC")
         for _ in range(num_dense_outer):
-            if dropout is not None:
+            if dropout is not None and dropout > 0:
                 self.outer_fc.add(layers.Dropout(float(dropout)))
 
             layer = layers.Dense(

bayesflow/networks/mlp/hidden_block.py

@@ -29,7 +29,7 @@ def __init__(
             self.dense = layers.SpectralNormalization(self.dense)
 
         if dropout is not None and dropout > 0.0:
-            self.dropout = layers.Dropout(dropout)
+            self.dropout = layers.Dropout(float(dropout))
         else:
             self.dropout = None
 

bayesflow/networks/mlp/mlp.py

@@ -60,7 +60,7 @@ def __init__(
         self.res_blocks.append(projector)
 
         if dropout is not None and dropout > 0.0:
-            self.res_blocks.append(layers.Dropout(dropout))
+            self.res_blocks.append(layers.Dropout(float(dropout)))
 
         for _ in range(depth):
             self.res_blocks.append(

bayesflow/networks/transformers/mab.py

@@ -59,7 +59,10 @@ def __init__(
                     use_bias=use_bias,
                 )
             )
-            self.feedforward.add(layers.Dropout(dropout))
+
+            if dropout is not None and dropout > 0:
+                self.feedforward.add(layers.Dropout(float(dropout)))
+
         self.feedforward.add(layers.Dense(output_dim))
         self.ln_post = layers.LayerNormalization() if layer_norm else None